Flowering ecology of a Box-Ironbark Eucalyptus community

Box-Ironbark forests occur on the inland hills of the Great Dividing Range in Australia, from western Victoria to southern Queensland. These dry, open forests are characteristically dominated by Eucalyptus species such as Red Ironbark E. tricarpa, Mugga Ironbark E. sideroxylon and Grey Box E. microcarpa. Within these forests, several Eucalyptus species are a major source of nectar for the blossom-feeding birds and marsupials that form a distinctive component of the fauna. In Victoria, approximately 83% of the original pre - European forests of the Box-Ironbark region have been cleared, and the remaining fragmented forests have been heavily exploited for gold and timber. This exploitation has lead to a change in the structure of these forests, from one dominated by large 80-100 cm diameter, widely -spaced trees to mostly small (≥40 cm DBH), more densely - spaced trees. This thesis examines the flowering ecology of seven Eucalyptus species within a Box-Ironbark community. These species are characteristic of Victorian Box-Ironbark forests; River Red Gum E. camaldulensis, Yellow Gum E. leucoxylon, Red Stringybark E. macrorhyncha, Yellow Box E. melliodora, Grey Box E. microcarpa, Red Box E. polyanthemos and Red Ironbark E. tricarpa. Specifically, the topics examined in this thesis are: (1) the floral character traits of species, and the extent to which these traits can be associated with syndromes of bird or insect pollination; (2) the timing, frequency, duration, intensity, and synchrony of flowering of populations and individual trees; (3) the factors that may explain variation in flowering patterns of individual trees through examination of the relationships between flowering and tree-specific factors of individually marked trees; (4) the influence of tree size on the flowering patterns of individually marked trees, and (5) the spatial and temporal distribution of the floral resources of a dominant species, E. tricarpa. The results are discussed in relation to the evolutionary processes that may have lead to the flowering patterns, and the likely effects of these flowering patterns on blossom-feeding fauna of the Box-Ironbark region. Flowering observations were made for approximately 100 individually marked trees for each species (a total of 754 trees). The flower cover of each tree was assessed at a mean interval of 22 (+ 0.6) days for three years; 1997, 1998 and 1999. The seven species of eucalypt each had characteristic flowering seasons, the timing of which was similar each year. In particular, the timing of peak flowering intensity was consistent between years. Other spatial and temporal aspects of flowering patterns for each species, including the percentage of trees that flowered, frequency of flowering, intensity of flowering and duration of flowering, displayed significant variation between years, between forest stands (sites) and between individual trees within sites. All seven species displayed similar trends in flowering phenology over the study, such that 1997 was a relatively 'poor' flowering year, 1998 a 'good' year and 1999 an 'average' year in this study area. The floral character traits and flowering seasons of the seven Eucalyptus species suggest that each species has traits that can be broadly associated with particular pollinator types. Differences between species in floral traits were most apparent between 'summer' and 'winter' flowering species. Winter - flowering species displayed pollination syndromes associated with bird pollination and summer -flowering species displayed syndromes more associated with insect pollination. Winter - flowering E. tricarpa and E. leucoxylon flowers, for example, were significantly larger, and contained significantly greater volumes of nectar, than those of the summer flowering species, such as E. camaldulensis and E. melliodom. An examination of environmental and tree-specific factors was undertaken to investigate relationships between flowering patterns of individually marked trees of E. microcarpa and E. tricarpa and a range of measures that may influence the observed patterns. A positive association with tree-size was the most consistent explanatory variable for variation between trees in the frequency and intensity of flowering. Competition from near-neighbours, tree health and the number of shrubs within the canopy area were also explanatory variables. The relationship between tree size and flowering phenology was further examined by using the marked trees of all seven species, selected to represent five size-classes. Larger trees (≥40 cm DBH) flowered more frequently, more intensely, and for a greater duration than smaller trees. Larger trees provide more abundant floral resources than smaller trees because they have more flowers per unit area of canopy, they have larger canopies in which more flowers can be supported, and they provide a greater abundance of floral resources over the duration of the flowering season. Heterogeneity in the distribution of floral resources was further highlighted by the study of flowering patterns of E. tricarpa at several spatial and temporal scales. A total of approximately 5,500 trees of different size classes were sampled for flower cover along transects in major forest blocks at each of five sample dates. The abundance of flowers varied between forest blocks, between transects and among tree size - classes. Nectar volumes in flowers of E. tricarpa were sampled. The volume of nectar varied significantly among flowers, between trees, and between forest stands. Mean nectar volume per flower was similar on each sample date. The study of large numbers of individual trees for each of seven species was useful in obtaining quantitative data on flowering patterns of species' populations and individual trees. The timing of flowering for a species is likely to be a result of evolutionary selective forces tempered by environmental conditions. The seven species' populations showed a similar pattern in the frequency and intensity of flowering between years (e.g. 1998 was a 'good' year for most species) suggesting that there is some underlying environmental influence acting on these aspects of flowering. For individual trees, the timing of flowering may be influenced by tree-specific factors that affect the ability of each tree to access soil moisture and nutrients. In turn, local weather patterns, edaphic and biotic associations are likely to influence the available soil moisture. The relationships between the timing of flowering and environmental conditions are likely to be complex. There was no evidence that competition for pollinators has a strong selective influence on the timing of flowering. However, as there is year-round flowering in this community, particular types of pollinators may be differentiated along a temporal gradient (e.g. insects in summer, birds in winter). This type of differentiation may have resulted in the co-evolution of floral traits and pollinator types, with flowers displaying adaptations that match the morphologies and energy requirements of the most abundant pollinators in any particular season. Spatial variation in flowering patterns was evident at several levels. This is likely to occur because of variation in climate, weather patterns, soil types, degrees of disturbance and biotic associations, which vary across the Box-Ironbark region. There was no consistency among sites between years in flowering patterns suggesting that factors affecting flowering at this level are complex. Blossom-feeding animals are confronted with a highly spatially and temporally patchy resource. This patchiness has been increased with human exploitation of these forests leading to a much greater abundance of small trees and fewer large trees. Blossom-feeding birds are likely to respond to this variation in different ways, depending upon diet-breadth, mobility and morphological and behavioural characteristics. Future conservation of the blossom-feeding fauna of Box-Ironbark forests would benefit from the retention of a greater number of large trees, the protection and enhancement of existing remnants, and revegetation with key species, such as E. leucoxylon, E. microcarpa and E. tricarpa. The selective clearing of summer flowering species, which occur on the more fertile areas, may have negatively affected the year-round abundance and distribution of floral resources. The unpredictability of the spatial distribution of flowering patches within the region means that all remnants are likely to be important foraging areas in some years.

Invertebrate larval dynamics in seasonally closed estuaries

Estuarine benthic assemblages are often numerically dominated by polychaetes. The limits of these populations are determined by larval, and probably to a lesser extent adult movement. A previous study (Newton 1996), indicated that planktonic polychaete larvae were very abundant over the summer months in the Hopkins River; however, the identification and source of these larvae was not known. Defining the extent of a population, and therefore the likelihood of that population recovering following a perturbation, is crucial for effective estuarine management. This study investigated both the likely source of the larvae, (i.e. estuarine or marine) and the extent of larval dispersal within and between estuaries by addressing the following questions: Which taxa produced the planktonic larvae? Are these taxa resident estuarine species? Are the larvae of different taxa evenly distributed within the estuary or do physicochemical parameters or other factors influence their abundance? Are the same larvae found in other estuaries along the coast? and Is there exchange of these larval taxa with the marine environment and other estuaries? Larvae were identified and described by culturing commonly occurring planktonic larvae until adult characteristics appeared. The spionids, Carazziella victoriensis and Prionospio Tatura, numerically dominated the plankton in the Hopkins and the spionid, Orthoprionospio cirriformia was recorded from the Hopkins, Curdies and Gellibrand estuaries. Two spionids, Carazziella sp. and Polydora sp. were identified from tidal waters. Mouth status and physicochemical conditions (salinity, temperature and dissolved oxygen) were monitored in each estuary. Whereas the Merri and Gellibrand estuaries were predominantly stratified over the sampling period, the Curdies was more often well mixed and the Hopkins varied from well mixed to stratified. The duration of mouth opening and hence the opportunity for larval exchange also varied in each estuary. The Merri River was closed for 13.5% of days over the study period, the Gellibrand River for 18.4%, the Hopkins River for 49% and the Curdies River for 71.0%. The distributions of larvae at spatial scales of metres, 100s of metres and kilometres were investigated within a single estuary. While the same larvae, C. victoriensis, P. Tatura and bivalve larvae, were found along the length of the Hopkins estuary the abundances varied at different spatial scales suggesting different processes were influencing the distribution of P. Tatura larvae, and C. victoriensis and bivalve larvae. The distribution of larvae between several estuaries was investigated by monitoring meroplankton at two sites at the mouth of each of the four estuaries approximately monthly (except for winter months). Different meroplanktonic assemblages were found to distinguish each estuary. Further, C. victoriensis and P. Tatura larvae were only recorded in the Hopkins but larvae of the spionid, Orthoprionopio cirriformia were detected in the Hopkins, Curdies and Gellibrand estuaries. The extent of larval exchange with other estuaries and the marine environment was determined by monitoring tidal waters. Settlement trays were also deployed to determine if larvae were moving into estuaries and settling but not recruiting. P. tatura larvae were not detected in the tidal waters of any estuary and while C. victoriensis and O. cirriformia were found in both flood and ebb tides there was no evidence of movement of theses taxa to other estuaries. Larvae of the spionids, Carazziella sp. and Polydora sp., were found in tidal waters of each estuary but were rarely detected in the plankton within the estuaries. Neither species was found as an adult in background cores from any estuary, nor with the exception of a few individuals in the Merri, were they detected in settlement trays in any estuary. I conclude that the source of the larvae of C. victoriensis, P. Tatura and O. cirriformia is estuarine and while C. victoriensis, and O. cirriformia move in and outh of the source estuary in tidal waters there was no evidence for movement to other estuaries. The spionids, Carazziella sp. and Polydora sp were considered to be marine and while they moved in and out of estuaries in tidal waters they did not usually settle in the estuaries. The results of this study are a crucial first step in the development of ecological models to better understand dispersal in seasonally closed estuaries that are typical of southern Australia. This study emphasises the unique physicochemical characteristics and biological assemblages within these estuaries and the need for estuarine management to reflect these differences.

Rearing juvenile Australian native percichthyid fish in fertilised earthen ponds

The post-larvae and fry of Australian native species, including those of species belonging to the family Percichthyidae, are routinely reared to a fingerling size (35-55 mm in length) in fertilised earthen fry rearing ponds. The juveniles of Murray cod (Maccullochella peelii peelii\ trout cod (Maccullochella macquariensis) and Macquarie perch (Macquaria australasicd) (Percichthyidae) are grown in fry rearing ponds at the Marine and Freshwater Resources Institute, Snobs Creek (Vie. Australia) for production of fingerlings for stock enhancement and aquaculture purposes. However, no detailed studies have been undertaken of the productivity of these ponds and factors that influence fish production. An ecologically based study was undertaken to increase the knowledge of pond ecology and dynamics, particularly in relation to the rearing of juvenile Murray cod, trout cod and Macquarie perch in ponds. Over nine consecutive seasons commencing in 1991, water chemistry, plankton, macrobenthos (2 seasons only) and fish were monitored and studied in five ponds located at Snobs Creek. A total of 80 pond fillings were undertaken during the study period. Additional data collected from another 24 pond fillings undertaken at Snobs Creek collected prior to this study were included in some analyses. Water chemistry parameters monitored in the ponds included, temperature, dissolved oxygen pH, ammonia, nitrite, nitrate, orthophosphate and alkalinity. Water chemistry varied spatially (within and between ponds) and temporally (diurnally, daily and seasonally). Liming of ponds increased the total alkalinity to levels that were considered to be suitable for enhancing plankton communities and fish production. Water quality within the ponds for the most part was suitable for the rearing of juvenile Murray cod, trout cod and Macquarie perch, as reflected in overall production (growth, survival and yield) from the ponds. However, at times some parameters reached levels which may have stressed fish and reduced growth and survival, in particular, low dissolved oxygen concentrations (minimum 1.18 mg/L), high temperatures (maximum 34°C), high pH (maximum 10.38) and high unionised ammonia (maximum 0.58 mg/L). Species belonging to 37 phytoplankton, 45 zooplankton and 17 chironomid taxa were identified from the ponds during the study. In addition, an extensive checklist of aquatic flora and fauna, recorded from aquaculture ponds in south-eastern Australia, was compiled. However, plankton and benthos samples were usually numerically dominated by a few species only. Rotifers (especially Filinia, Brachionus, Polyarthra, and Asplanchnd), cladocerans (Moina and Daphnid) and copepods (Mesocyclops and Boeckelld) were most abundant and common in the plankton, while chironomids (Chironomus, Polypedilum, and Prodadius) and oligochaetes were most common and abundant in the benthos. Both abundance and species composition of the plankton and macrobenthos varied spatially (within and between ponds) and temporally (diurnally, daily and seasonally). Chlorophyll a concentrations, which ranged from 1.8 to 184 \ig/L (mean 29.37 ng/L), initially peaked within two weeks of filling and fertilising the ponds. Zooplankton peaked in abundance 2-4 weeks after filling the ponds. The maximum zooplankton density recorded in the ponds was 6,621 ind./L (mean 721 ind./L). Typically, amongst the zooplankton, rotifers were first to develop high densities (2nd-3rd week after filling), followed by cladocerans (2nd-4th week after filling) then copepods (2nd-5th week after filling). Chironomid abundance on average peaked later (during the 5th week after filling). The maximum chironomid density recorded in the ponds was 27,470 ind./m2 (mean 4,379 ind./m2). Length-weight, age-weight and age-length relationships were determined for juvenile Murray cod, trout cod and Macquarie perch reared in ponds. These relationships were most similar for Murray cod and trout cod, which are more closely related phylogenetically and similar morphologically than Macquarie perch. Growth of fish was negatively correlated with both size at stocking and stocking biomass. Stocking density experiments showed that, at higher densities, growth offish was significantly reduced, but survival was not affected. The diets of juvenile Murray cod trout cod and Macquarie perch reared in fry ponds were similar. The cladocerans Moina and Daphnia, adult calanoid and cyclopoid copepods and the chironomids, Chironomus, Polypedilum and Procladius were the most commonly occurring and abundant prey. Selection for rotifers and copepod nauplii was strongly negative for all three species of fish. Size range of prey consumed was positively correlated with fish size for trout cod and Macquarie perch, but not for Murray cod. Diet composition changed as the fish grew. Early after stocking the fish into the ponds, Moina was generally the more common prey consumed, while in latter weeks, copepods and chironomids became more abundant and common in the diet. On a dry weight basis, chironomid larvae were the most important component in the diets of these fish species. Selective feeding by fish on larger planktonic species such as adult copepods and cladocerans, may have influenced the plankton community structure as proposed by the trophic cascade or top -down hypothesis. The proximate composition and energy content of Murray cod, trout cod and Macquarie perch, reared in the ponds did not vary significantly between the species, and few significant changes were observed as the fish grew. These results suggested that the nutrient requirements of these species might not vary over the size range of fish examined. Significant differences in the proximate composition of prey were observed between species, size of species and time of season. The energy content of prey (cladocerans, copepods and chironomids) on a pond basis, was closely related to the abundance of these taxa in the ponds. Data collected from all pond fillings during the present study, along with historical data from pond fillings undertaken prior to this study, were combined in a data matrix and analysed for interactions between pairs of parameters. In particular, interactions between selected water chemistry parameters, zooplankton and chironomid abundance indicators were analysed to identify key factors that influence fish production (growth, survival, condition and yield). Significant correlations were detected between fish production indicators and several water chemistry and biota (zooplankton and chironomids) parameters. However, these were not consistent across all three species of fish. These results indicated that the interactions between water chemistry, biota and fish were complex, and that combinations of these parameters, along with other factors not included in the present study, may influence fish production in these ponds. The present study, showed that more stringent monitoring of fry rearing ponds, especially water quality, zooplankton and benthos communities and fish, combined with an associated increase in understanding of the pond ecosystem, can lead to substantial improvements in pond productivity and associated fish production. In the present study this has resulted in a general increase in fish survival rates, which became less variable or more predictable in nature. The value of such knowledge can provide managers with a more predicative capacity to estimate production of ponds in support of stock enhancement programs and provision of juvenile for aquaculture grow-out.

Multilabel classification by BCH code and random forests

This paper uses error correcting codes for multilabel classification. BCH code and random forests learner are used to form the proposed method. Thus, the advantage of the error-correcting properties of BCH is merged with the good performance of the random forests learner to enhance the multilabel classification results. Three experiments are conducted on three common benchmark datasets. The results are compared against those of several exiting approaches. The proposed method does well against its counterparts for the three datasets of varying characteristics.

Underestimated and severe : small mammal decline from the forests of south-eastern Australia since European settlement, as revealed by a top-order predator

In Australia, numerous small mammal species have suffered extinction or severe declines in distribution and abundance following European settlement. The extent of these declines from forested areas of south-eastern Australia, however, remains poorly understood. In this paper we use sub-fossil deposits of the sooty owl (Tyto tenebricosa tenebricosa) as a tool for understanding the diversity of the small mammal palaeocommunity. These results are compared to the contemporary sooty owl diet from the same geographical region to investigate the degree of small mammal decline following European settlement. Of 28 mammal species detected in sub-fossil deposits and considered prey items of the sooty owl at the time of European settlement, only 10 species were detected in the contemporary sooty owl diet. Numerous small mammal species have not only recently suffered severe declines in distribution and abundance but have also recently undergone niche contraction, as they occupied a greater diversity of regions and habitats at the time of European settlement. For some species our understanding of their true ecological niche and ecological potential is therefore limited. The species that underwent the greatest declines occupied open habitat types or were terrestrial. The severity of decline is also likely to have resulted in severe disruption of ecosystem functions, with wide scale ecosystem consequences. There is an urgent need to improve small mammal conservation, to maintain crucial ecosystem functions performed by small mammals. It is recommended that broad-scale exotic predator control programs are conducted which may also provide suitable conditions for the re-introduction of locally extinct species.

Chemical composition and tissue energy density of the cuttlefish (Sepia apama) and its assimilation efficiency by Diomedea albatrosses.

The cuttlefish Sepia apama Gray (Mollusca: Cephalopoda) is a seasonally abundant food resource exploited annually by moulting albatrosses throughout winter and early spring in the coastal waters of New South Wales, Australia. To assess its nutritional value as albatross forage, we analysed S. apama for water, lipid protein, ash contents, energy density and amino acid composition. Because albatrosses consistently consume S. apama parts preferentially in the order of head, viscera and mantle, we analysed these sections separately, but did not identify any nutritional basis for this selective feeding behaviour. The gross energy value of S. apama bodies was 20.9 kJ/g dry mass, but their high water content (>83%; cf <70% for fish) results in a relatively low energy density of 3.53 kJ/g. This may contribute to a need to take large meals, which subsequently degrade flight performance. Protein content was typically >75% dry mass, whereas fat content was only about 1%. Albatrosses feed on many species of cephalopods and teleost fish, and we found the amino acid composition of S. apama to be comparable to a range of species within these taxa. We used S. apama exclusively in feeding trials to estimate the energy assimilation efficiency for Diomedea albatrosses. We estimated their nitrogen-corrected apparent energy assimilation efficiency for consuming this prey to be 81.82 ± 0.72% and nitrogen retention as 2.90 ± 0.11 g N kg^-1 d^-1. Although S. apama has a high water content and relatively low energy density, its protein composition is otherwise comparable to other albatross prey species. Consequently, the large size and seasonal abundance of this prey should ensure that albatrosses remain replete and adequately nourished on this forage while undergoing moult.

Compensatory growth in an aquatic plant mediates exploitative competition between seasonally tied herbivores

The degree to which vertebrate herbivores exploitatively compete for the same food plant may depend on the level of compensatory plant growth. Such compensation is higher when there is reduced density-dependent competition in plants after herbivore damage. Whether there is relief from competition may largely be determined by the life-history stage of plants under herbivory. Such stage-specific compensation may apply to seasonal herbivory on the clonal aquatic plant sago pondweed (Potamogeton pectinatus L.). It winters in sediments of shallow lakes as tubers that are foraged upon by Bewick's Swans (Cygnus columbianus bewickii Yarrell), whereas aboveground biomass in summer is mostly consumed by ducks, coots, and Mute Swans. Here, tuber predation may be compensated due to diminished negative density dependence in the next growth season. However, we expected lower compensation to summer herbivory by waterfowl and fish as density of aboveground biomass in summer is closely related to photosynthetic carbon fixation. In a factorial exclosure study we simultaneously investigated (1) the effect of summer herbivory on aboveground biomass and autumn tuber biomass and (2) the effect of tuber predation in autumn on aboveground biomass and tuber biomass a year later. Summer herbivory strongly influenced belowground tuber biomass in autumn, limiting food availability to Bewick's Swans. In contrast, tuber predation in autumn by Bewick's Swans had a limited and variable effect on P. pectinatus biomass in the following growth season. Whereas relief from negative density dependence largely eliminates effects of belowground herbivory by swans, aboveground herbivory in summer limits both above- and belowground plant biomass. Hence, there was an asymmetry in exploitative competition, with herbivores in summer reducing food availability for belowground herbivores in autumn, but not the other way around.

Spatial ecology of sooty owls in south-eastern Australian coastal forests : implications for forest management and reserve design

We investigated the home-range size and habitat use of eight Sooty Owls (Tyto tenebricosa tenebricosa) in coastal forests in East Gippsland, Victoria, Australia, between November 2006 and January 2008. The size of home-ranges varied widely; based on 95% adaptive kernel estimates, the average size of home-ranges of males was 3025ha (±1194s.d., n=3), whereas that of females was 994ha (±654s.d., n=5). Sooty Owls utilised a broad range of ecological vegetation classes and topographical features for roosting and foraging at a greater scale than previously assumed. There was minimal selection for habitat types based on floristic composition, primarily only avoiding heathlands (for foraging and roosting) and selecting particular dense foliage (rainforest and riparian scrub) for foliage roosting. Two Owls maintained home-ranges close to logged areas, with logging regrowth (<45 years old) being strongly avoided by both individuals. We recommend that the size of individual reserves for Sooty Owls in commercial forests should be increased to more closely resemble the core spatial resource requirements needed by a pair. Reserves should be largest where they feed predominantly on hollow-dependent prey. Most importantly, rather than conservation measures just focussing on the spatial requirements of Sooty Owls, efforts should be directed towards retaining high densities of crucial resources, such as hollow-bearing trees and mammalian prey species throughout the landscape.